On Jul 7, Tony Laszlo wrote:
> Jochen and all,
>
> Please let me know if my sending such a lengthy report
> is not appropriate for these fora.
Its better to have too much information, than too few.
> Thank you very much for your time and assistance.
> Conclusions first: still no success with Trakker.
>
> This specs for this machine are as follows:
>
> I have also been trying various bios and jumper adjustments.
> After my initial failures with ftape and Trakker, I tried
> Trakker under Win95. I found that, while the device worked
> on a separate Windows machine, it didn't work on the Linux
> service that I need it on (temporarily running under Win95).
> It seemed that the parallel port LPT1 was faulty. Now, this
> port is on the motherboard together with several other ports
> including the one that goes to the display, mouse, etc.
> I could not find any jumper which might disable the
> faulty parallel port.
>
> IBM Aptiva 520 2144-68J (DX486CPU/ 100Mhz, 60M Ram, ISA bus)
> So, I bought a new ISA multiport card which as a parallel
> port. I think this is it being recognized by dmesg:
>
> Serial driver version 4.27 with HUB-6 MANY_PORTS MULTIPORT SHARE_IRQ
> DETECT_IRQ enabled
>
> * Not sure if it means anything, but occasionally (not always) I get
> this in dmesg:
>
> Jul 7 00:06:55 net kernel: Floppy drive(s): fd0 is 1.44M
> Jul 7 00:06:55 net kernel: DOR0=4
> Jul 7 00:06:55 net kernel: floppy interrupt on bizarre fdc 1
> Jul 7 00:06:55 net kernel: handler=00000000
> Jul 7 00:06:55 net kernel: FDC 0 is a post-1991 82077
You shouldn't use IRQ 6 for the parallel port, since it is normally
used by the floppy controller. It is better to use IRQ 5 for the
second parallel port.
> I then took a look at the bios setting for the machine.
> There is only one setting for parallel port. It had been
> set to LPT 1 at 0x378 IRQ=7 and SPP. The only other options are
> LPT 1 at 0x3BC IRQ=7, LPT 2 at 0x278 IRQ=7 and Disabled.
> (Port modes are ecp & bi-di, epp & bi-di, ps/2 & bi-di
> and compatible [maybe that is SPP]).
For maximum performance you should use a bi-di mode, e.g. ps/2, for
the trakker. But any mode should work.
> None of this has had much of an effect on what happens
> when I do insmod parport and insmod parport_pc.
It is not easy to autodetect the irq number for parallel ports. AFAIK
only ECP makes it possible to detect the interrupt. You should always
use the "irq=" option of parport_pc.
> Here's a sample of what's in the /var/log/messages after
> insmod trakker.
>
> trakker.c: trakker_register @ 0xc485def0
> [002] fdc-io.c (fdc_register_R7d7204fe) - Probing for trakker tape drive slot 0.
> [003] 0 fdc-parport.h (ft_parport_probe) - dev: c3886760.
> [004] 0 fdc-parport.h (ft_parport_probe) - irq: -1.
> [005] 0 fdc-parport.h (ft_parport_probe) - port: 378.
This means, that for the first parallel the IRQ was not set nor
detected. Therefore the port is just skipped by the trakker driver.
> [006] 0 fdc-parport.h (ft_parport_probe) - dev: c3886760.
> [007] 0 fdc-parport.h (ft_parport_probe) - irq: 5.
> [008] 0 fdc-parport.h (ft_parport_probe) - port: 278.
> [009] 0 trakker.c (trakker_checksum) - checksum error (off by 63).
> [010] 0 trakker.c (trakker_checksum) - checksum error (off by 63).
It seems that there is no trakker connected to the second parallel
port at 0x278, or it is not turned on. The irq setting doesn't matter
here, only the port address. The numbers "off by 63" and "off by eb"
are clear signs that nothing is responding/connected.
Another hint: Sometimes it is necessary to "hard reset" the drive.
Just remove the power plug for a few seconds. You don't need to have
a tape inserted for your tests.
Jochen
-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2b29 : Fri Jul 07 2000 - 04:21:23 EDT